Robotic Pancreaticoduodenectomy for Pancreatic Head Tumour: A Single-Centre Analysis

Background: The robotic approach is an appealing way to perform minimally invasive pancreaticoduodenectomy. We compare robotic cases' short-term and oncological outcomes to a historical cohort of open cases. Methods: Data were collected in a prospective database between 2016 and 2024; complications were graded using the ISGPS definition for the specific pancreas-related complications and the Clavien-Dindo classification for overall complications. Furthermore, the Comprehensive Complication Index was calculated. All patients undergoing pancreaticoduodenectomy were included, except those with acute or chronic pancreatitis, vascular tumour involvement or multi-visceral resections. Only the subset of patients with malignancy was regarded for the oncologic outcome. Results: In total, 100 robotic and 102 open pancreaticoduodenectomy cases are included. Equal proportions of patients have a main pancreatic duct <= 3 mm (p = 1.00) and soft consistency of the pancreatic remnant (p = 0.78). Surgical time is longer for robotic pancreaticoduodenectomy (p < 0.01), and more patients have delayed gastric emptying (44% and 28.4%, p = 0.03). In the robotic group, the number of patients without any postoperative complications is higher (p = 0.02), and there is less chyle leak (p < 0.01). Ninety-day mortality, postoperative pancreatic fistula, and postpancreatectomy haemorrhage are similar. The lymph node retrieval and R0 resection rates are comparable. Conclusions: In conclusion, after robotic pancreaticoduodenectomy, remembering all cases during the learning curve are included, less chyle leak is observed, the proportion of patients without any complication is significantly larger, the surgical duration is longer, and more patients have delayed gastric emptying. Oncological results, i.e., lymph node yield and R0 resection rate, are comparable to open pancreaticoduodenectomy.